272 1'ETTEBSSOK, OK WATER AND ICE. 



the regular expansion of the solid hydrates HCOOH and 

 CH3COOH for many degrees below their melting point. The 

 considerable change of volume accompanying the melting pro- 

 cess of these compounds then begins at a premature stage 

 of temperatnre and gradually increases, until the whole mäss 

 is transformed into the liquid state. 



In the same degree that the acetic acid was purified from 

 adhering traces of water, these signs of premature melting 

 rapidly diminished, and the cnrve of regular exj^ansion of 

 volume began to extend to the vicinity of the real melting 

 point, but nevertheless the purest hydrate, which I could 

 prepare by repeated distillation and crystallisation from 10 

 kilogrammes of ordinary acid. acet. puriss. did not show a 

 sudden transition, at any definite temperatnre, of the specific 

 volume of the solid into that of the liquid substance. I there- 

 fore suspected, that a slight trace of foreign substances would 

 cause a similar anomaly in the melting-process of the ice, to 

 that, which a trace of water occasions in the behavior of the 

 solid acetic or formic hydrate; and that the negative results, 

 hitherto obtained by Pl ii c k er and Geissler as well as b}- 

 myself, were due to the absolute purity of the water. I resolved 

 first to try ordinary distilled water from one of the glass reser- 

 voirs of the laboratory. The water had been kept there more 

 than a week, protected from dust etc, but in communication 

 "vvith the air of the laboratory room. Silver nitrate and chlorid 

 of mercur}^ with carbonate of sodium occasioned a faint opalisa- 

 tion in the fiuid; other agencies were powerless. A dröp of 

 the water, cautiously evaporated on a glass plate, left a visible 

 residue. I concluded that, wuth the exception of slight traces 

 of ammonium salts and of chlorine, the water was pure. The 

 volumes of ice from this sample of water are recorded in table 

 III. In the vicinity of the melting point the ice shows a re- 

 markable contraction of volume, ' which seems to begin already 



1 This fact recalls to our mind the old hypothesis, once supported '\>\ 

 Muschenbroek and de Mairan, that the ice expands its volume, when 

 cooled, and contracts, on being heated. Although this opinion was refuted 

 by the experiments of Placidus Heinrich already in the beginning of 

 this century, it was revived by Petzholdt in 1843, who tried to explain the 

 movement of the glaciers by the dilatation of the ice, caused by the winter 

 cold, and its contraction by the heat of the summer. 



In order to corroborate bis theory, Petzholdt determined the expan- 

 sion of ice by weighing a silver bottle filled with pure frozen water in ether 

 at different temperatures (from — 2° to — 8 W) below zero. In fact the 

 coefficient of expansion was found to be negative at all temperatures. 



